There have been many efforts in recent years to explain differences in the performance of commercial banks. Interest has centered on the extent to which changes in a selected group of indices of bank performance are related to tile structure of bankinq markets and selected other factors thought to influence bank behavior. While various tecnniques have been used, the most common has been multiple linear regression. The measures of performance entered into the regression equations have included the price and quantity of bank services and bank profitability, while the explanatory variables have included, to name only a few, the one-, two-, or three-bank concentration ratio, the number of banks in the market, the existence of competition from nonbank financial institutions, bank costs, bank size, and proxies for the demand for banking services. Generalizations then have been made about the impact of market structure and other variables on bank performance, generalizations based upon the regression coefficients of the explanatory variables. The consensus appears to be that the demand for banking services and bank costs are significant determinants of the performance of individual commercial banks; market structure appears to be much less important. However, the conclusions are by no means unanimous. Numerous conceptual and data problems are common to these studies. Not the least of these limitations is the isolated treatment of each of the performance variables. In fact, the performance of commercial banks should not be measured by a single proxy variable, such as the loan rate or bank profits, but by a set of variables which are jointly determined by market structure, demand, and other factors. The specification of the model should not rest upon a single measure of bank performance (as in the regression studies cited above) but should be multivariate (for example, a set of simultaneous equations). The purpose of this paper is to report the results of the... [ABSTRACT FROM AUTHOR]